1. Field of the Invention
The present invention is generally related to testing and determining the quality and coherence of a laser beam for use in photolithography systems.
2. Related Art
The ability to fabricate integrated circuit chips with increasingly smaller feature sizes depends upon continual evolution of photolithographic methods. Typically, a light source is used to illuminate a mask (reticle) so that a pattern is transferred into photoresist applied to an underlying semiconductor wafer. Machines that performs this operation are referred to as wafer steppers or wafer scanners. In order to achieve an accurate representation of the reticle pattern at submicron dimensions on the photoresist, it is necessary to use a light source that can support both a high degree of resolution and depth of focus. This requirement has led to the use of lasers as light sources for photolithographic applications.
However, the use of laser light for photolithography is not without its drawbacks. The high degree of coherence in the light produced by a laser gives rise to situations whereby interference among rays within the beam can produce a random distribution of the intensity of the light within a cross section of the beam. This random distribution of light intensity is known as speckle. Speckle adversely affects the development of the photoresist and therefore has been the subject of a myriad of corrective efforts. As speckle is an unwanted by product of the coherent property of laser light, the ability to measure coherence is a useful first step in correcting for speckle.
Coherence of a beam of light occurs when the rays within the beam travel parallel to one another and their corresponding wavefronts remain in phase over time. The extent to which these qualities are achieved is referred to as the degree of coherence. Often coherence is viewed as having two components: temporal (or longitudinal) coherence and spatial coherence. Temporal coherence measures deviations in frequency about a nominal frequency. Spatial coherence is a measure of how collimated a beam is. If a beam is highly collimated, the phases of its wavefronts are nearly identical at a given cross section of the beam.
Interference is a phenomenon that occurs when coherent beams of light overlap or intersect. Waves of light consist of oscillating fields of electric and magnetic energy. When beams of light overlap or intersect, the intensity of the light at the points of intersection is a function of the interaction among the fields of electric and magnetic energy at those points. The nature of this interaction depends upon the degree of coherence of the intersecting beams. Where the intersecting beams have a high degree of coherence, the intensity of the light at the points of intersection is proportional to the square of the vector sum of the amplitudes of the fields of electric and magnetic energy. However, if the intersecting beams are highly incoherent, the intensity of the light at the points of intersection is proportional to the sum of the square of the amplitudes of the fields of electric and magnetic energy. Therefore, if coherent beams are substantially in phase at the points of intersection, the intensity of the light is greater than the contribution of each individual beam. The points of intersection appear brighter than their surroundings. This is referred to as constructive interference. However, if coherent beams are significantly out of phase at the points of intersection, the intensity of the light is lesser than the contribution of each individual beam. The points of intersection appear dimmer than their surroundings. This is referred to as destructive interference.
As interference is a phenomenon produced by the interaction of coherent beams of light, analysis of an interference pattern created when two portions of a coherent beam of light are made to interfere with each other can be used to measure the degree of coherence. Typically, the degree of coherence is expressed as a coherence length, relating to the distance of separation, in time or space, between the two portions of the coherent beam of light creating the interference pattern. Coherence length has traditionally been measured using interferometers. Interferometers operate by splitting a coherent beam of light into two portions and later recombining the two portions to observe the resulting interference pattern. To test for temporal (longitudinal) coherence, the path length of one of the portions is extended to impart a delay in time. For spatial coherence, each portion is extracted from a separate area within the cross section of the beam. While measuring the intensity of the constructive interference areas within the interference pattern, the distance of separation is increased until the intensity falls below a specific figure of merit. The distance of separation at this point is the coherence length. The figure of merit is usually given as a percentage of the maximum intensity measured, but other figures of merit can also be used. Typical cutoff percentages are based on exponential decay or points where intensity or power are half of their maximum measured values.
Classic designs of interferometers include the Michelson, the Fabry-Perot, and the Fizeau. These are well known in the art. These instruments make use of movable arrangements of beam splitters, mirrors, and half-silvered mirrors to manipulate the paths of the beams. Much effort in the art has been expended to improve these basic designs. Ironically, where in photolithography it is desirable to reduce coherence, efforts to develop a high quality interferometer based on the classic designs seek the ability to measure coherence in real time so that it can be increased for use in high quality interferometer calibration. Where lasers are used for photolithography, the classic designs have several disadvantages: (1) the susceptibility of the instrument to inaccuracies arising from vibrations induced not only by moving parts, but also by the introduction of purge gases that, depending upon the wavelength of the light, may be needed to minimize absorption along the optical paths; (2) the difficulty of controlling the precise position of moving parts of the instrument; (3) the possibility that disassembly of optical train parts can change preset alignments; (4) the inherently fragile nature of the design; (5) the complexity involved in fabricating parts for the instrument; and (6) the expense incurred in manufacturing a sensitive instrument.
What is needed is an instrument that: (1) is insensitive to vibrations; (2) has no moving parts; (3) minimizes the extent of disassembly of optical train parts; (4) is inherently robust in design; (5) is simple to manufacture; and (6) is inexpensive. What is also needed is an instrument that can readily support real time measurement of coherence so that be increased for use in high quality interferometer calibration.